Impact of window function on statistical index of quality of spectral estimates of poliharmonic acoustic signals

As the role of passive acoustic systems of extracting information is increasing in military sphere the interest to algorithms of polyharmonic acoustic signals detection has grown greatly. The difficulty of detecting such signals is caused by a priory uncertainty of its structure. Using statistical index of quality (SIQ) of spectral estimates for adaptive thresholding simplify implementing this task. The paper gives the expressions for SIQ computation of periodogram and its overall values are defined for cases of harmonic oscillation and white noise. It is proved that SIQ pulsation level (difference between its' maximum and minimum values) is connected to pulsation level of overall amplitude spectrum of Fast Fourier Transform. Ascertained that SIQ pulsation level decreases with spreading main lobe of window function in frequency domain. We proposed to use for weighting polyharmonic acoustic signals Tukey, Cauchy or Dolph?Chebyshev windows, which simultaneously provide minimal SIQ pulsation level and coefficient of signals' energy losses.